Male circumcision to cut HIV risk in the general population

In today’s Lancet, results from two randomised trials show considerable benefit of male circumcision in reducing HIV incidence in men, confirming findings from an earlier trial in South Africa. The recent focus on antiretroviral treatment—although necessary and appropriate—has made primary prevention a secondary issue. Renewed interest in HIV prevention is welcome. The South African trial was done in a periurban setting near Johannesburg, in 3274 men aged 18–24 years randomised to immediate (n=1617) or later (n=1657) circumcision. The trial was stopped after a planned interim analysis showed a significant 60% relative reduction in HIV risk associated with circumcision. The participants were from the general population and loss to follow-up was low, supporting the generalisability of the findings. However, concern was expressed about the randomisation procedures, the slight imbalance in baseline characteristics between groups, and potential selection bias. WHO and other UN agencies issued a supportive statement, but urged restraint while awaiting the results of two ongoing trials, presented today. Robert Bailey and colleagues’ trial in Kisumu, Kenya, was also done in men aged 18–24 years, randomly assigned to circumcision (n=1391) or delayed circumcision (n=1393). The trial was stopped after an unscheduled interim analysis in December, 2006, when evidence emerged of a significant benefit from circumcision. The HIV incidence rate was high, at 1·1 per 100 person-years in the circumcised men and 2·1 per 100 person-years in the control group (or 2·1% and 4·2%, respectively, over 2 years). There was an estimated 53% (unadjusted modified intention-to-treat analysis) to 60% (as-treated analysis) reduction in relative risk of HIV infection associated with male circumcision. Ronald Gray and co-workers’ study in Rakai, Uganda, included 4996 men aged 15–49 years; 2474 were randomised to immediate and 2522 to delayed circumcision. As with the trial in Kenya, this trial was stopped early after an interim analysis showed significant efficacy. HIV incidence was 0·66 per 100 person-years in the circumcision group and 1·33 per 100 person-years in the control group. The estimated reduction in the relative risk of infection with HIV was 51% (unadjusted modified intention-to-treat analysis) to 55% (as-treated analysis). Both trials were methodologically and analytically sound. Thus three randomised trials now provide firm evidence that the risk of acquiring HIV is halved by male circumcision. Applying the circumcision efficacy to the age-specific HIV incidence rates from a large population-based longitudinal HIV survey in rural KwaZulu-Natal, South Africa, we find that circumcision would prevent an estimated 35 000 new HIV infections in 2007 alone in the 2·5 million mostly uncircumcised men in the province (figure 1).